Post-randomization bias

Cook, Chad; Garcia, Alessandra Narciso

doi:10.1080/10669817.2020.1739153

Cited by 8 publications

(2 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Possible reasons for changes in outcomes reported may include the loss of funding, data collection delays, failing to achieve an adequate sample size, or deciding after registration to report an outcome frequently collected by clinicians 149 . Regardless of intention, outcomes not aligned with original methodology, an issue in both RCTs and SRs, 24,28,149,150 can be misleading 149,151 …”

Section: Discussionmentioning

confidence: 99%

“…149 Regardless of intention, outcomes not aligned with original methodology, an issue in both RCTs and SRs, 24,28,149,150 can be misleading. 149,151 The registration rate of SRs vary widely among different disciplines, with registration rates ranging from 3% to 43.9%. [152][153][154][155] In a PROSPERO-specific analysis, a random sample of 500 SRs indexed in PubMed from 2011 to 2018 found only 15% of eligible reviews had been registered in PROSPERO.…”

Section: Missing Itemsmentioning

confidence: 99%

See 1 more Smart Citation

Trials and tribulations of transparency related to inconsistencies between plan and conduct in peer‐reviewed physiotherapy publications: A methodology review

McClenahan

Lojacono

Young

et al. 2023

Evaluation Clinical Practice

View full text Add to dashboard Cite

RationaleThe physiotherapy profession strives to be a leader in providing quality care and strongly recognizes the value of research to guide clinical practice. Adherence to guidelines for research reporting and conduct is a significant step towards high‐quality, transparent and reproducible research.Aim/ObjectiveAssess integrity between planned and conducted methodology in randomized controlled trials (RCTs) and systematic reviews (SRs) published in physiotherapy journals.MethodsEighteen journals were manually searched for RCTs and SRs published from 1 July 2021 through 31 December 2021. Studies were included if the journal or specific study was indexed in PubMed and published/translated in English. Descriptive statistics determined congruence between preregistration data and publication.ResultsForty RCTs and 68 SRs were assessed. Forty‐three SRs included meta‐analysis (MA). Of the 34 registered RCTs, 7 (20.6%) had no discrepancy between the registration and publication. Two trials (5.9%) addressed all discrepancies, 4 (11.8%) addressed some and 21 (61.8%) did not address any discrepancies. Of the 36 registered MAs, 33 (91.7%) had discrepancies between the registration and publication. Two (5.6%) addressed all discrepancies and three (8.3%) had no discrepancies. Eight SRs without MA published information not matching their registration, and none provided justification for the discrepancies.ConclusionMost RCTs/SRs were registered; the majority had discrepancies between preregistration and publication, potentially influencing the outcomes and interpretations of findings. Journals should require preregistration and compare the submission with the registration information when assessing publication suitability. Readers should be aware of these inconsistencies and their implications when interpreting and translating results into practice.

show abstract

Section: Discussionmentioning

confidence: 99%